Re: Suggested response to the Yandex "cannot iive with loosening of TAG participation requiremens"

> On Apr 16, 2015, at 8:12 , Daniel Glazman <daniel.glazman@disruptive-innovations.com> wrote:
> 
> On 15/04/15 15:19, David Singer wrote:
> 
>> As I said, I and many others have had actual experience of counter-examples to (1), and there is plenty of research indicating how strong (2) is and how difficult to resist.
>> 
>> There are other reasons that people have put forward, earlier in the discussion, and you have not addressed them either.
> 
> 
> I can't address a problem I don't see. We have two entirely different
> perspectives: you assume an issue is probable and take 'a priori'
> counter-measures that I find too heavy. I assume the issue is highly
> unprobable or at least easily contained and I prefer 'a posteriori'
> measures to 'a priori' counter-measures, and you find that too light.

To be clear, I think we can relax from the current counter-measure (duplicate representation cannot occur at all) to a position where we allow it until the next regularly scheduled election.  So, I am not proposing adding counter-measures at all, but relaxing ones we already have.

I find the arguments for relaxing beyond the next election not as strong as those for relaxing up to the next election. Indeed, this process change was focused on the narrow case of whether we can improve the handling of a change of affiliation. I think any change to allowing full-term serving after a change of affiliation would only make sense in the context of a wholesale removal of the rule against having more than one representative from the same organization, which is a change with a much larger scope.

> I trust Tim and TAG Co-chairs to shut down any attempt by a vendor to
> impose undesirable views to the whole TAG. You prefer representativity limitations.

If we know we disagree with a position, I am also confident of people’s ability to resist it. I am less confident of our ability to notice when we are being swayed, when ‘group think’ is taking us into directions we might not otherwise espouse. I am also concerned that in a small group, having two from the same employer may reduce the diversity of opinion that would be expressed at all (the existence of a second person from the same employer means that we may have less diversity of view.)

> These are two irreconciliable views, I guess.

To an extent, yes.

In general, processes do seem to be a pain when things are going well. No-one needs processes when all is sweetness and honey and we agree, Processes help guide us so we don’t end up in nasty tar-pits (a-priori measures, as you say).  (It’s not relevant here, but they also provide vehicles to help manage situations where all is not going well.) The trick with the preventive aspects of a process is to have them prevent the bad while not impacting the good; as usual, we never quite succeed in preventing all bad, and we never quite succeed in having no impact to the good. We try to find a favorable balance.

David Singer
Manager, Software Standards, Apple Inc.